Despite the long-standing literature that has demonstrated changes in driving capacity following brain injury (BI) ? little is known about the relationship of these differences and increased risk for driver error or the prediction of long term driving outcome after BI. However, it is well-established that the loss of the driving privilege negatively impacts functional re-integration, mood and quality of life ? resulting from the reduced ability to participate in various life activities, work and educational experiences. The challenge to increasing our understanding of how to best assess and predict driving performance after BI is two-fold. First, there is a need for novel assessment methodologies that can provide objective, detailed and repeatable metrics of driving performance. The current clinical gold standard ? the behind the wheel (BTW) driving assessment, is over-dependent on subjective observations, lacks standardization, assesses only basic driving skills (due to safety limitation) and generates gross measures of performance (i.e., Pass/Fail). Second, there is a lack of follow-up studies that examine actual return to driving behaviors among individuals with BI. While some evidence for greater risk of crash involvement (often dichotomized as Yes/No) has been reported, these studies have relied heavily on self-reported data and offer little to no data about driver behaviors and/or modifications, risk-involvement, crash causing-behaviors or driving patterns. The proposed study aims to address these limitations and employs an established virtual reality driving simulator (VRDS) that outputs novel driving performance metrics that are currently not available thru clinical methodology. The VRDS generates detailed metrics that can differentiate between clinical populations. Specifically, the study will integrate VRDS into an existing clinical driving assessment program and evaluate 100 individuals with BI across the process of returning to drive (e.g., from assessment to follow- up) and a sample of healthy controls. All participants will be assessed with both current clinical protocols and VRDS. This will be followed by a 24 month follow-up study including an innovative, 3-platform approach (in- car video-monitoring, web-based self-report and driving records) to quantifying returned to driving behaviors. The data collected will be used to apply both traditional (Regression Models) and novel (Machine-Leaning Models) analytical techniques to generate predictive models of relevant outcome variables (i.e., risk involvement, crash-relevant errors) that can be used to inform tailored driver interventions and retraining.